A comparative study on the availability of modern contraceptives in public and private health facilities in a peri-urban community in Ghana

Study design

A cross-sectional design was used allowing for a comparative analytical approach.
A mixed-methods approach involving quantitative and qualitative data collection techniques
was used.

Study area

Ghana is divided into districts, municipalities and metropolis Data for the study
was collected in the Ga East municipality, a predominantly peri-urban community located
in the Greater Accra region of Ghana 16]. The rapidly developing population makes it ideal to study. The municipality is made
up of four sub-municipalities with a total population of 320,853 and over 100 health
facilities ranging from hospitals and clinics to pharmacies and licensed chemical
shops 16].

Sampling methods and data collection

The health facilities in the municipality served as the sampling frame for the study.
For the purposes of the comparative analytical approach, the health facilities were
first divided into two broad arms: private and public health institutions. These two
broad arms were made up of four major types of health facilities depending on the
level of care each facility provided. These four types of health facilities were the
only available health facilities where clients could obtain family planning commodities
in the district. The four types were: hospitals, clinics//maternity homes/health care
centres, pharmacies and licensed chemical shops. A sample of 51 facilities from an
estimated 150 health facilities were used. These were made up of 43 private and 8
public health facilities.

The municipality had only eight public health facilities where family planning services
were rendered and all were used for this study. For the private institutions, a total
of three different lists covering all the private health facilities in the municipality
was employed to select the final sample. The first list comprising all maternity homes/clinics/
health centres and hospitals in the municipality where family planning services were
provided was obtained from the municipality’s health directorate in Abokobi, the capital
of the municipality. Health facilities where family planning services were not provided
were excluded from the list.

The municipality’s health directorate could not provide a clear cut list for pharmacies
and chemical shops. However, they were instrumental in the acquisition of two representative
lists covering all pharmacies and chemical shops. These were obtained from the government
facilities in the four sub-municipalities under the municipality.

The three different lists showed that pharmacies provided the most number of private
health facilities with hospitals providing the least. The forty three (43) private
facilities used were distributed as follows: pharmacies 20 (45 %), chemical shops
13 (35 %), clinics/polyclinics/health centres/maternity homes 7 (15 %) and hospitals
3 (5 %). This was in accordance with the proportion each category of health facility
contributed.

Simple random sampling was employed to select from each of the three lists the required
number of facility under each category. With the help of the community health nurses
in the respective sub-municipalities, each of the randomly selected facilities under
the various categories was located and the availability of modern contraceptives in
that facility determined using a checklist. We adopted and adapted a previously used
checklist in a related study 17]. The current study’s checklist, which was pre-tested in a few selected health facilities
outside the study area, had four main sections with corresponding questions ranging
from the age and sex of the respondent to the name and location of the health facility.
The checklist defined availability of a contraceptive as follows:

For a contraceptive to be available, the facility manager first of all indicated that
the facility had that contraceptive in stock as in Bowen et al.17]. Additionally, the current study impressed upon facility managers to provide evidence;
and this was observed by the trained data collection personnel. Finally, for each
type of contraceptive, the trained personnel ensured that the quantity in stock met
the minimum contraceptive stock level for that facility.

The facility manager present in each selected facility was also interviewed by a trained
personnel and the interview recorded. The in-depth interview guide used by the trained
personnel also had four main sections covering key indicators such as factors affecting
availability and cost of modern contraceptives in the facility. A total of 51 in-depth
interviews (IDIS) were conducted amongst the facility managers.

Power calculations

A descriptive study by Bowen et al.17] provided information on the availability of male condoms and IUDs in the municipality
and this was used in the calculating the sample size. The availability of male condoms
in the government health facilities was 71 % whilst that of the private health facilities
was 25 %. Given the lack of data on the other modern contraceptives, this was used
as a proxy. Using a 95 % confidence level and a statistical power of 80 %, the minimum
sample size of 48 was obtained 18] as illustrated below. At the end of the study however, 51 facilities participated.

Sample size calculations (illustration)

.

Data analysis and management

A comparative assessment of the two groups, private and public health facilities was
done after data entry and cleaning. We used univariate analysis to generate descriptive
tabulations for key variables. Statistical tools such as frequency distribution tables
and cross tabulation were used to compare the availability of modern contraceptives
in public and primary health facilities. Analysis looking at the associations between
certain attributes of the facilities (type, place, and average number of clients per
day) and the main outcome variable, availability of each category of contraceptive
identified, was tested using logistic regression. Associations were assessed using
multiple logistic regression techniques, where odds ratios (ORs) and their 95 % confidence
intervals (CI) were computed. We employed a standard logistic regression modelling
in SPSS (the “Enter” method) in our analysis. With this method, all the variables
previously reported to be associated with the outcome variable or found to be associated
with the outcome during the bivariate analysis were entered and a full model generated
in a single step. A key attribute of each model (R
²
) are included in Table 2. P value 0.05 was used to denote statistical significance. All analyses were performed
using IBM SPSS Statistics for Windows, Version 20.0.

The key outcome variables assessed in this study was fertility intentions/childbearing
desires (the desire to a child or children in the near future), awareness and use
of contraceptives (being aware of and usage of various contraceptive options available
to HIV-positive women).

For the qualitative aspect, the running notes from the in-depth-interviews were written
into expanded notes. The audio recordings were also transcribed and read several times
to clean the data. The information collected was listed in Microsoft word 2010 and
then categorized manually into emerging themes. The themes were not determined prior
to the start of the study.

Ethical approval and informed consent

Ethical clearance was obtained from the Ghana Health ervice Ethical Review Committee
of the Research and Development Division of the Ghana Health Service. In addition
to the ethical clearance, permission was also sought from the Municipal director of
the Ga East municipal health directorate, as well as the facility managers of the
various private and public health facilities.

Informed consent was sought from the facility managers of the various health facilities
and it included: the purpose of the study, the risks and benefits, privacy and confidentiality
as well as conflict of interest.